IREN Limited (IREN) Earnings

IREN Limited is expected to report next earnings on November 5, 2026 (in NaN days), with a consensus EPS estimate of $-0.50. IREN has beaten EPS estimates in 3 of its last 12 reported quarters (average surprise -172.7% over the last four).

Next earnings
Nov 5, 2026in NaN days
EPS est $-0.50 · Revenue est $258M
Track record
Beat EPS in 3 of 12 quarters
Avg surprise -172.7% (last 4 quarters)
Earnings history
Report dateEPS estEPS actualSurpriseRevenueRev. surprise
Aug 27, 2026$-0.55$-1.89-243.1%$137M-4.1%
May 7, 2026$-0.18$-0.25-38.9%$145M-36.1%
Feb 5, 2026$-0.24$-0.44-82.0%$121M-48.5%
Nov 6, 2025$0.15$-0.34-326.8%$240M+1.5%
Aug 28, 2025$0.18$0.08-54.7%$187M-0.9%
May 14, 2025$0.17$0.11-35.3%$145M-15.8%
Feb 12, 2025$-0.04$0.09+325.0%$118M-30.0%
Nov 26, 2024$-0.07$-0.22-214.3%$54M-2.2%
Aug 28, 2024$-0.02$-0.27-1105.4%$57M-1.9%
May 15, 2024$0.05$0.07+40.0%$54M+4.8%
Feb 15, 2024$0.03$-0.07-358.1%$42M+8.6%
Sep 13, 2023$-0.17$-0.12+27.9%$34M+0.2%

Source: company filings + earnings calendar. For informational purposes only — not investment advice.

Earnings call summary

Q4 FY2026 · August 27, 2026

AI summary of management’s prepared remarks and analyst Q&A. For informational purposes only — not investment advice.

Management highlights

- **Strategic Positioning**: IRNF operates across three layers—data centers/power, compute (GPUs), and software (Mirantis)—owning the full stack to maximize value at each layer. This vertical integration allows them to offer managed services and enterprise support alongside bare-metal compute. - **Customer Growth & Contracts**: Secured new multi-year contracts with major clients including Cohere, Prometheus, Perplexity, Figure AI, and a leading undisclosed frontier AI lab. Existing customers like Together AI and Fireworks AI have renewed and expanded. - **Infrastructure Delivery**: Delivered 'Horizon 1' (50 MW liquid-cooled deployment) to Microsoft, achieving NVIDIA Exemplar Cloud status. Horizons 2-4 are targeted for December delivery. Total 2026 IT load target is ~300 MW; 2027 target is an additional 500 MW. - **Global Expansion**: Active construction sites in Texas, British Columbia, Oklahoma, South Australia, and Spain. Decided to convert Canal Flats (BC) to liquid cooling for GB300s to leverage spare power headroom. - **Financing Success**: Raised $6.5 billion in GPU financing over the past three months. Prepayments from customers fund 45-55% of GPU CapEx. Non-investment grade financing closed at single-digit rates (e.g., 9% fixed rate for McKinsey deployments). - **Operational Efficiency**: Design templates (Horizon series) are being reused to reduce build times. Leveraging spare power capacity at existing sites to add more GPUs without new grid connections.

Guidance

- **FY2027 CapEx**: Guarding capital expenditure of approximately $25–$30 billion for the fiscal year ending June 2027. This covers 2026 contracted capacity, air-cooled deployments in calendar 2027, and early-stage liquid-cooled capacity for late 2027. - **Revenue Ramp**: Expects ARR to exceed $4 billion by the end of the December quarter. Significant revenue recognition is expected to occur predominantly in the March quarter due to late-quarter capacity activations. - **Funding Plan**: Targeting an additional $8 billion in GPU financing and prepayments to support GPU CapEx requirements within the $25–$30 billion capex envelope. The remainder will be met through data center financing, operating cash flows, and corporate sources. - **Mining Decommissioning**: Expects mining operations to be effectively decommissioned by the end of December 2026.

Segment performance

The provided transcript does not break down financial performance by specific product segment (Data Center, Compute, Software). However, it provides aggregate financial data: Total revenue for the June quarter was $137.2 million, with AI cloud revenue contributing $70.5 million. The company reported a net loss of $684 million for the quarter, primarily driven by non-cash impairments related to decommissioning mining hardware. Contracted Annual Recurring Revenue (ARR) stands at $4 billion for 2026 capacity, with $1 billion currently operating.

Risks & headwinds

- **Transition Costs**: Significant non-cash impairments ($450.4 million) were recorded due to the decommissioning of mining hardware, highlighting the operational risk and cost associated with pivoting infrastructure from crypto-mining to AI compute. - **Execution Risk**: Bringing GPUs online is identified as the primary bottleneck, not signing contracts. Delays in construction, supply chain, or GPU delivery could impact revenue realization. - **Capital Intensity**: The business requires massive ongoing capital expenditure ($25–$30 billion annually). Failure to secure sufficient debt or equity financing, or rising interest rates, could constrain growth. - **Market Dependency**: While diversifying, the company remains dependent on the continued high demand for AI infrastructure. A slowdown in AI adoption or shifts in customer spending could affect utilization and pricing.

Analyst Q&A

  • Q: Goldman Sachs asked about Mirantis's role in commercialization and whether it enabled signed deals or could sell to aggregators like OpenRouter. /

    A: CCO Kent Draper explained Mirantis opens new customer classes (smaller labs/enterprises needing orchestration) and enables on-demand compute. CEO Daniel Roberts added that Mirantis is now an NVIDIA-certified hypervisor, enhancing their full-stack value proposition. They do not explicitly plan to use Mirantis for aggregator sales but view it as a way to expand addressable market and attach higher-margin services.

  • Q: Goldman Sachs followed up on the $25-30B FY27 CapEx guidance, asking if it supports only 800MW of 2027 contracts or includes future years, and requested details on funding the balance beyond the $14B secured. /

    A: CFO Anthony Lewis clarified the CapEx covers 2026 deliveries, all 2027 air-cooled deployments, and H2 2027 liquid-cooled builds (excluding next year's GPU costs for those). He noted $14B is secured, with another ~$8B estimated from GPU financing/prepayments. The rest will come from unencumbered data center asset finance and operating cash flows, leveraging strong fundraising conditions.

  • Q: Macquarie asked about the decision to pivot British Columbia sites to liquid cooling, questioning if it was driven by organic demand or speculative excess power returns. /

    A: Kent Draper stated the pivot is driven by actual customer demand for liquid-cooled GPUs and effective design for BC conditions. Paul Goding then asked about allocating excess liquidity toward DC build vs. speculative compute purchases. Draper confirmed they are doing both, prioritizing DC build to satisfy immediate scarcity while adding compute/software layers for better economics. Dan Roberts added that prepayments often cover 100% of DC costs, creating a self-funding flywheel.

  • Q: Canaccord asked if the financing environment can sustainably support industry-wide build-out pace, expressing concern about capital availability. /

    A: CFO Anthony Lewis cited evolving markets for both GPU and DC financing, with participation from private/public markets and investment-grade lenders. CEO Dan Roberts argued financing follows real demand, citing how GPU financing formed in months unlike real estate's decade. He emphasized that short payback periods (~2 years) and upfront prepayments make AI infra highly attractive to lenders like PIMCO/Blue Owl, ensuring capital availability as long as demand persists.